a b s t r a c t

Objective: Although timely administration of antibiotics has an established benefit in serious bacterial infection, the majority of studies evaluating antibiotic delay focus only on the first dose. Recent evidence suggests that delays in redosing may also be associated with worse clinical outcome. In light of the increasing burden of boarding in Emergency Departments (ED) and subsequent need to redose antibiotic in the ED, we examined the association between delayed second antibiotic dose administration and mortality among patients admitted from the ED with a broad array of infections and characterized risk factors associated with delayed second dose administration.

Methods: We performed a retrospective cohort study of patients admitted through five EDs in a single healthcare system from 1/2018 through 12/2018. Our study included all patients, aged 18 years or older, who received two intravenous antibiotic doses within a 30-h period, with the first dose administered in the ED. Patients with end stage renal disease, cirrhosis and extremes of weight were excluded due to a lack of consensus on antibiotic dos- ing intervals for these populations. Delay was defined as administration of the second dose at a time-point greater than 125% of the recommended interval. The primary outcome was in-hospital mortality.

Results: A total of 5605 second antibiotic doses, occurring during 4904 visits, met study criteria. Delayed admin- istration of the second dose occurred during 21.1% of visits. After adjustment for patient characteristics, delayed second dose administration was associated with increased odds of in-hospital mortality (OR 1.50, 95%CI 1.05-2.13). Regarding risk factors for delay, every one-hour increase in allowable compliance time was associ- ated with a 18% decrease in odds of delay (OR 0.82 95%CI 0.75-0.88). Other risk factors for delay included ED boarding more than 4 h (OR 1.47, 95%CI 1.27-1.71) or a high acuity presentation as defined by emergency sever- ity index (ESI) (OR 1.54, 95%CI 1.30-1.81 for ESI 1-2 versus 3-5).

Conclusions: Delays in second antibiotic dose administration were frequent in the ED and early hospital course, and were associated with increased odds of in-hospital mortality. Several risk factors associated with delays in second dose administration, including ED boarding, were identified.

(C) 2021

1. Introduction

? Presented at: Annual Meeting of the Society of Academic Emergency Medicine, Las Vegas NV, May 2019.

* Corresponding author.

E-mail addresses: [email protected] (C.B. Kemmler), [email protected] (R.B. Sangal), [email protected] (C. Rothenberg),

[email protected] (S.-X. Li), [email protected] (F.S. Shofer), [email protected] (B.S. Abella), [email protected] (A.K. Venkatesh), [email protected] (S.D. Foster).

Timely antibiotic administration has an established role in quality sepsis care [1,2], as well as in the care of other infections, such as neutro- penic fever [3,4], bacteremia [5] and meningitis [6]. While the majority of clinical guidelines, quality metrics and research to date have focused on the time to initial antibiotic administration, little attention has been directed towards the relative importance of timely administration of subsequent doses. In an era of increased ED boarding and frequent care-team transitions, adherence to appropriate repeat antibiotic dosing can be challenging.

https://doi.org/10.1016/j.ajem.2021.02.058

0735-6757/(C) 2021

Abbreviations

ED emergency department ESI emergency severity index EMR electronic medical record

CCI Charlson Comorbidity Index CrCl creatinine clearance

ICU intensive care unit

Control of bacterial infections is dependent on the maintenance of sufficient pharmacokinetic and pharmacodynamic parameters, suggest- ing an important role for the appropriate timing of antibiotic doses be- yond the Initial administration in achieving optimal clinical outcomes [7-9]. Despite this, only two prior studies have examined the association of clinical outcomes, including mortality, with delays in second antibi- otic dose administration [10,11]. In a single-center study by Leisman et al., an exploratory analysis of clinical outcomes in septic patients showed that delayed second doses were associated with increased odds of in-hospital mortality [10]. In a subsequent study of 90 patients with severe sepsis or septic shock, Andersson et al. observed an in- creased odds of mortality in a subpopulation of patients who experi- enced a delay in first or second antibiotic dose administration, however a significant association was not seen in patients who experi- enced a delay in first or second dose alone [11]. It is also notable that both of these previous studies assessed the role of delay in only septic patients.

Accordingly, we sought to utilize a large, diverse, multi-center dataset to assess for an association between delays in second antibiotic dose administration in the treatment of a broad spectrum of infectious indications and in-hospital mortality, as well as to explore potential risk factors for delays.

1. Methods
   1. Study design

We conducted a retrospective study of ED patient encounters lead- ing to hospital admission across five hospitals within a single healthcare system between January 1, 2018 and December 31, 2018. Admitted pa- tients who received two or more doses of the same intravenous antibi- otic within a 30-h period, the first of which was administered in the ED, were included in the study. Study data were abstracted from the healthcare system electronic medical record (EMR), which was com- mon across all study sites. The study was reviewed and deemed exempt by the institutional review board.

• 1. Study setting and population

The five hospitals in this study included two academic and three community facilities, in a mix of urban and suburban locations. Eligible patients were 18 years of age or older and received at least two or more intravenous doses of any one of a defined set of antibiotics for a defined indication in a 30-h period, with the first dose occurring in the ED (Sup- plemental Table 1). Upon placing an antibiotic order within our health system, clinicians are prompted to identify the indication for antibiotics within the medication order. Eligible study indications included severe sepsis, neutropenic fever and infections of the respiratory tract, urinary tract, abdomen/pelvis, bone/joint, central nervous system and skin and soft tissue. These were selected based on the presence of common anti- biotic redosing guidelines across our health system. The 30-h redosing window was chosen to ensure sufficient detection of episodes of de- layed administration of antibiotics with recommended administration frequencies of 24 h, but to also limit the detection of instances of initially discontinued antibiotics that were later restarted for a different

indication (ex. hospital-acquired infection). Exclusion criteria included end stage renal disease, dialysis dependence, cirrhosis, body mass index greater than 40 kg/m², and weight greater than 120 kg or less than 40 kg. These exclusion criteria were recommended by institution ED pharmacists, as they impart altered pharmacokinetic and pharmaco- dynamic qualities of the study antibiotics, precluding the use of the standardized dosing frequency algorithms [8,12,13]. ED visits for which there was missing EMR data were also excluded (Fig. 1).

• 1. Study protocol

Patient demographic data were obtained from the EMR, including age, gender, weight, Emergency Severity Index category, initial creatinine, and medical comorbidities. Time points collected included: first and second antibiotic dose order, first and second antibiotic dose administration, inpatient bed order, and patient transfer from ED. Addi- tional clinical data were obtained, including antibiotic order indication, admitting service, and in-hospital mortality. Charlson Comorbidity Index score was based on previously published literature [14]. ED boarding was defined as a patient being present in the ED for more than 4 h after a hospital bed order was placed in EMR [15].

Creatinine clearance (CrCl) was calculated for each patient using the Cockroft-Gault formula. Study antibiotics fulfilling the following criteria were chosen for study inclusion: intravenous formulation and the abil- ity to determine a recommended dosing frequency based only on pa- tient CrCl and indication for antibiotic. Appropriate antibiotic dosing frequencies, based on weight, CrCl and disease indication, were deter- mined based on protocols developed by the institutional Antimicrobial stewardship Program (Supplemental Table 2) [16-18]. The interval be- tween the initial administration of an antibiotic and the administration of the second dose was calculated, and the administration of the second dose was deemed ‘non-compliant’ if it was administered at a time inter- val greater than 125% of the recommended dosing interval, indicating a delay of 25% of more [10,11]. For the analysis of ED visits in which the patient received more than one antibiotic, lack of compliance for any one of the antibiotics designated the visit as noncompliant for subse- quent analyses of associations with mortality and factors delaying anti- biotic administration.

• 1. Outcome measures

The primary outcome of interest was in-hospital mortality. Second- ary outcomes were the association of delayed second antibiotic dose ad- ministration with potential risk factors, including gender, age, CCI, ESI, recommended antibiotic dosing frequency, ED boarding status, patient location at time of recommended second dose administration (pre- or post-transfer from ED), admitting service and the use of separate orders for first and second antibiotic doses.

• 1. Data analysis

We utilized a standard query to obtain demographics, timestamps and aforementioned variables of interest. Logistic regression was per- formed for the primary outcome of in-hospital mortality adjusting for age, gender, CCI, ESI, boarding, creatinine, antibiotic delay, admission to Intensive Care Unit (ICU)-level care. A similar logistic regression was performed for the outcome of antibiotic delay. Data analysis was conducted using R version 3.6.3 [19].

• 1. Sensitivity analysis

As a sensitivity analysis, regression models (as stated in 2.5 Data Analysis) were repeated with the inclusion of lactate values, among pa- tients for whom this data was available, to explore whether findings were robust to an additional marker of clinical severity. Given that de- fining noncompliant second dose delivery as one occurring after 125%

Fig. 1. Emergency department visit identification and inclusion in the study. A query of the EMR identified 6015 events of second antibiotic dose administration meeting study inclusion and exclusion criteria. 410 of these events were excluded due to missing study data. The resulting 5605 events were used for study analysis and occurred during a total of 4904 ED visits and associated admissions. 4488 unique patients were treated at these 4904 visits.

of the recommended dosing interval provides a more permissive timeframe for antibiotics with longer recommended frequencies, we re- peated our analysis with a more explicit definition of delay, defining noncompliant delivery as the redosing of any study antibiotic at a time-point greater than two hours after the recommended time for sec- ond dose administration. This analysis is reflective of a standardized ap- proach to timely medication delivery often used in real-world bedside care and reasonable for medications with a 24-h recommended fre- quency, such as some of the study antibiotics [20,21]. Lastly, in recogni- tion of potential overlap in antimicrobial coverage when multiple antibiotics were used, a third sensitivity analysis was performed in which regression models were repeated using an alternative outcome definition in which compliance for visits utilizing more than one study antibiotic was dictated by compliance of the antibiotic with the shortest dosing frequency.

1. Results
   1. Study population

We identified 6015 events meeting the inclusion criteria (Fig. 1). Of these, 410 (6.8%) were excluded due to lack of data necessary to calcu- late the recommended redosing interval. We included the remaining 5605 events in the analysis. These events occurred during 4904 ED visits, 672 (13.7%) of which entailed the administrations of two or three eligible antibiotics. The mean patient age for ED visits was

61.6 years, and patients were male in 49.7% of visits (Table 1). Indica- tions for antibiotics included, among all of those evaluated in this study (Supplemental Fig. 1), severe sepsis/septic shock (13.9% of or- ders) and infections of the lower respiratory tract (35%), abdomen/pel- vis (23.5%), urinary tract (13.4%) and central nervous system (2.6%).

Table 1

ED visit characteristics.

Characteristic	Total	2nd antibiotic dose compliant	2nd antibiotic dose noncompliant	p value
Total visits	4904	3870	1034
Patient age, mean (SD)	61.6 (19)	62.7 (19)	57.8 (18)	<0.0001
Patient gender, male	2435 (50%)	1891 (49%)	544 (53%)	0.035
Charlson Comorbidity Index = 0	413 (8%)	335 (9%)	78 (8%)	0.351
Charlson Comorbidity Index = 1	341 (7%)	262 (7%)	79 (8%)
Charlson Comorbidity Index >= 2	4150 (85%)	3273 (85%)	877 (85%)
ESI = 1	92 (2%)	65 (2%)	27 (3%)	<0.0001
ESI = 2	2722 (56%)	2047 (53%)	675 (65%)
ESI = 3	2056 (42%)	1730 (45%)	326 (32%)
ESI = 4	32 (1%)	26 (1%)	6 (1%)
ESI = 5	2 (0%)	2 (0%)	0 (0%)
Patient boarding (>=4 h)	2141 (44%)	1600 (41%)	541 (52%)	<0.0001
Admission to Internal Medicine	2814 (57%)	2368 (61%)	446 (43%)	<0.0001
Admission to General Surgery	282 (6%)	223 (6%)	59 (6%)
Admission to Medicine subspecialty	1451 (30%)	1060 (27%)	391 (38%)
Admission to Surgery subspecialty	357 (7%)	219 (6%)	138 (13%)
Total events	5605	4552	1053
Use of antibiotic with 4-h recommended frequency?	24 (0%)	6 (0%)	18 (2%)	<0.0001
Use of antibiotic with 6-h recommended frequency?	1229 (22%)	602 (13%)	627 (60%)
Use of antibiotic with 8-h recommended frequency?	1042 (19%)	792 (17%)	250 (24%)
Use of antibiotic with 12-h recommended frequency?	1091 (19%)	933 (20%)	158 (15%)
Use of antibiotic with 24-h recommended frequency?	2219 (40%)	2219 (49%)	0 (0%)
Single standing order used for administration of 1st and 2nd antibiotic doses?	821 (15%)	683 (15%)	138 (13%)	0.4096

Data is based on ED visit unless otherwise noted. Listed percentages are proportions of column totals. A chi square test was used for all categorical analyses (ex. gender), and t-tests were used for all continuous number analyses (ex. age).

* Indicates data based on instances of use of a study antibiotic instead of ED visit.

Table 2

Delayed antibiotic redosing is associated with a significant increase in in-hospital mortal- ity. Regression was performed for antibiotic delay alone, boarding alone and both antibi- otic delay and boarding together. A sensitivity analysis was performed to additionally control for lactate level, although exclusion of 23% of the study sample was required due to missing lactate data for those visits.

Mortality Outcome OR Lower CI Upper CI p-value Antibiotic delay (no boarding) 1.51 1.06 2.15 0.022?

Boarding (no antibiotic delay) 1.14 0.82 1.59 0.44

with increased mortality. The significantly increased odds of mortality associated with delayed second antibiotic dose administration persisted when both ED boarding and antibiotic delay were included as covariates in the same statistical analysis (Antibiotic delay OR 1.50, 95% CI 1.05-2.13, p = 0.024).

• 1. Secondary outcomes

Combined Model

_* Analysis of short antibiotic redosing frequencies as a risk factor for

Abx delay 1.50 1.05 2.13 0.024

Boarding 1.12 0.80 1.56 0.50

delayed second dose administration revealed that the frequency of compliance increased stepwise as the recommended dosing interval in-

Sensitvity analysis with lactate added to model (loss of 23% of sample)

_* creased, ranging from 25% for a 4-h interval to 100% for a recommended

Antibiotic delay (no boarding) 1.47 1.01 2.11 0.038

Boarding (no antibiotic delay) 1.21 0.86 1.71 0.26

Combined Model

Abx delay	1.47	1.01	2.10	0.041?
Boarding	1.21	0.86	1.70	0.28

Models controlled for age, gender, creatinine, CCI, ED to ICU, and ESI. Includes secondary covariates only (excludes mortality and ICU variables).

* Significant association.

Of these 4904 ED visits, 1034 (21.1%) involved a delay in a second antibiotic dose administration greater than 25% of the recommended frequency (Table 1). Full summary data are shown in Table 1.

• 1. Primary outcome

Given that prior literature has suggested ED boarding may contrib- ute to poor outcomes [22-25], we performed multiple regression models to understand whether antibiotic delay or ED boarding may ex- plain the clinical outcome findings (Table 2). With regards to mortality, antibiotic delay alone (OR 1.51, 95% CI 1.06-2.15, p = 0.022) and not ED boarding alone (OR 1.14, 95% CI 0.82-1.59, p = 0.44) was associated

24-h interval (Table 1 and Fig. 2).

The evaluation of additional potential risk factors for delayed sec- ond dose administration is shown in (Fig. 3). Some factors associated with a significantly increased odds of noncompliance included an ED boarder status (OR 1.47, 95% CI 1.27-1.71, p < 0.0001), a high acuity presentation (OR 1.54, 95% CI 1.30-1.81, p < 0.0001 for ESI 1-2 ver- sus 3-5), multiple medical comorbidities (OR 1.84, 95% CI 1.36-2.51, p < 0.001 for a CCI of 2 or greater), and younger age (OR 0.98, 95% CI 0.98-0.99, p < 0.0001). Factors not associated with an increased odds of noncompliance included gender and arrival creatinine level. Antibiotic dosing was not included in regression analyses given that patients can receive multiple antibiotics. However, as shown in Fig. 2, shorter antibiotic redosing intervals were noncom- pliant more frequently and, in an unadjusted model, every one- hour increase in permissible time for compliant administration was associated with an 18% decrease in odds of delay (OR 0.82, 95% CI 0.75-0.88, p < 0.0001).

Sixteen percent (N = 932) of second antibiotic doses in the study were due prior to patient transfer from the ED to inpatient care. Assessing for a potential association between patient location at the time of a recommended second dose administration and compliance,

Fig. 2. Distribution of actual antibiotic interval administration by recommended frequency. Median and interquartile range shown for each dosing frequency. Percentage of second doses that were provided within the permissible timeframe for each recommended dosing frequency are shown on the right. Every one-hour increase in permissible time for compliant administration was associated with an 18% decrease in odds of delay (OR 0.82 95% CI 0.75-0.88, p < 0.0001).

Fig. 3. Odds ratio of delay for various characteristics in all study patients. Reference groups for categorical variables are Gender = female, Lactate = no, CCI = 0, ESI = 3-5, Service = General Medicine, Boarding <=4 h, ICU Floor = no. *p < 0.05, **p < 0.001.

we found a similar proportion of second doses were noncompliant when recommended administration was while the patient was in the ED as compared to when the patient had already transferred from the ED (51.3% vs 50.9%, respectively, p = 0.973) (Table 3). To evaluate whether a recommended second dose delivery time around the time of patient transfer from the ED might be associated with poor com- pliance, we found that a similar proportion of administrations were noncompliant when they were due within the 1 h before or 1 h after transfer as compared to when they were due outside of the time imme- diately before or after transfer (Table 3).

• 1. Sensitivity analysis

In our sensitivity analysis also controlling for lactate level, and subsequently limited to the 3805 (77.6%) visits with lactate results

available, we found similar results. Antibiotic delay alone was asso- ciated with mortality (OR 1.47, 95% CI 1.01-2.11, p = 0.038) and ED boarding alone was not associated with mortality (OR 1.21, 95% CI 0.86-1.71, p = 0.26). In the model including both antibiotic delay and ED boarding, antibiotic delay was also associated with mortality (OR 1.47, 95% CI 1.01-2.10, p = 0.041).

In a sensitivity analysis using a more explicit definition of noncom- pliant delay, defining it as the delay of second dose administration by greater than 2 h for any of the study antibiotics, we also observed similar results to our primary analysis (Supplemental Tables 3 and 4). Finally, in a third sensitivity analysis examining potential overlap of coverage of the target organism by multiple antibiotics used in the same patient, in which compliance for a visit was dictated by compliance of the anti- biotic with the shortest dosing frequency, similar results were also dem- onstrated (data not shown).

Table 3

Compliance of second antibiotic administration does not vary based on the relative time of patient transfer from ED. Standardizing to second antibiotic doses with a recommended 6 h frequency, the frequency of noncompliant second doses was determined for doses that were due prior to 1 h before patient departure from the ED, within 1 h before patient departure from the ED, within 1 h after patient departure from the ED, and later than 1 h after patient departure from the ED.

Time of recommended second antibiotic dose relative to time of patient transfer from ED	>1 h before patient left ED	<=1 h before patient left ED	<=1 h after patient left ED	>1 h after patient left ED
Total visits during timeframe	243	28	22	936
Number of noncompliant second doses	124 (51.0%)	15 (53.6%)	11 (50.0%)	477 (51.0%)

Number of noncompliant second doses 139 (51.3%)? 488 (50.9%)

* (51.3% vs 50.9%, p = 0.973).

1. Discussion

Although the role of early antibiotics in the narrow infectious spec- trum of sepsis is well established, the large majority of studies to date have focused on the time to first antibiotic administration. We found that this important relationship between antibiotic timing and bacterial suppression may extend, not only to the essential timeliness of repeat antibiotic doses, but also to a broader population of ED patients with Acute infections. Specifically, we report an elevated risk of mortality among patients in whom the second antibiotic dose was delayed. This finding is consistent with the findings of Leisman et al., in one of only two prior studies to previously assess clinical outcomes associated with a delay in second antibiotic administration in patients admitted from the ED, and extends these findings to a larger cohort of patients with a more diverse set of indications for antibiotic administration [10]. Furthermore, the Leisman et al. study was exploratory in nature, and we support those findings here with a hypothesis-driven, confirma- tory analysis.

Evaluating potential risk factors for delay of second antibiotic dose administration, we found an association, similar to that seen in prior work, between shorter dosing frequency and risk of delay [10]. We also observed a significant independent association between antibiotic delay and ED boarding. Prior studies have shown that ED operations are not traditionally optimized for extendED patient care times [26,27]. In fact, prior literature has suggested a relationship between ED boarding and mortality [22-25]. While we did not find the same associ- ation in this analysis, ED boarding was a risk factor for antibiotic delay (Fig. 3). We hypothesize that the correlation between delay and ED boarding status may be related to a lack of sufficient ED practices to en- sure the ordering of repeat doses or the lack of established ED nursing practices for monitoring the need for repeat medication doses. Further, ED boarding might also cause delays in antibiotic administrations due after patient admission, as handoffs to admitting teams are often not given by the primary ED team, and they potentially involve the need for a greater volume of information due to longer patient care times.

Interestingly, an association was also seen between a high acuity presentation (ESI 1 or 2) and risk of delayed second dose administra- tion. Although this is somewhat counter-intuitive, given the expectation that more acutely ill patients will get more attentive care, potential ex- planations may include a higher burden of other interventions by the care team and a greater complexity of patient handoffs, in which appro- priate timing of repeat doses may not be sufficiently communicated.

Other interesting findings include an observed association between

elevated creatinine level and compliant second dose administration, as well as one between increased age and compliant second dose adminis- tration. The association with creatinine level might relate to the role of CrCl in calculating appropriate antibiotic frequency. An elevated creati- nine increases the likelihood that antibiotic timing will need to be ad- justed, and subsequent communication of that adjustment from a pharmacist to the ED or inpatient care teams might serve a protective role as a reminder for timely second dose administration in patients with acute or chronic kidney disease. An independent association be- tween advanced age and compliant dosing might relate to the poten- tially increased attentiveness of clinical staff to older individuals or perhaps greater participation of older patients in their care due to, on average, more experience with hospital care.

Our analysis of the relationship between ED to inpatient transitions and delay of second antibiotic dose administration yielded unexpected findings. Chiefly, we anticipated that second doses due around the time of transition would have a higher odds of delay, which was not demonstrated by our analysis. Similarly, we were surprised to find that the use of standing antibiotic orders with specified redosing inter- vals were not more effective in promoting compliance than one-time orders in the ED, suggesting that the resolution of delays in second dose administration in the ED will require more than simple standardi- zation to the use of a standing antibiotic order for all doses.

These findings carry important clinical quality and policy implica- tions. First, clinical quality improvement programs should broaden their focus beyond initial antibiotic administration to also measure and improve the timeliness of second dose administration. Within our healthcare system, this might include the utilization of standardized handoff strategies both at ED shift change and during patient handoff to admitting teams [28,29]. And, as the use of standing antibiotic orders did not significantly improve timely antibiotic redosing in our study, standardized nursing handoffs should also be considered [30]. Given the intensity of ED and institutional focus on sepsis care in recent years, existing dashboards and Quality Improvement teams should be able to flexibly expand existing efforts to ensure better acute care. Sec- ond, policymakers should consider the development of quality mea- sures to promote improvements in adherence to antibiotic dosing regimens. Our study demonstrates that data elements required for mea- surement are all feasibly captured within existing electronic health re- cords and several studies have now documented both a quality gap as well as evidence of association between this tangible process of care and outcomes.

1. Limitations

There are several important limitations to the present study. First, the retrospective nature of the study risks incorporation of bias and confounders into our study groups, as may be evidenced by the uneven distribution of several demographic characteristics across our study groups. We attempted to address these limitations through the use of regression analyses. And as with all retrospective study findings, the increased odds of mortality found in this study was only associated with delayed second antibiotic doses, and not necessarily caused by them. As such, delays in antibiotic delivery may serve as a correlate for other deficiencies in care that are re- sponsible for this outcome. Additionally, we recognize the limita- tions of data abstraction, and in this case specifically the possibility that timestamps collected from the electronic medical records may not reflect the exact time that antibiotic doses were given. We be- lieve the large sample size likely minimizes the likelihood that errors in documentation influenced our results in one direction.

We would also like to note that by only including patients in which two doses of the same antibiotic were given within 30 h of one another, we may have inadvertently excluded instances of non- compliant administration of a second dose of antibiotics with a 24-h recommended frequency. Closer examination of the data set for an- tibiotics with a 24-h recommended frequency demonstrated that a general downtrend exists in the number of patients represented in hours 24 through 30 following initial dosing and only 18 patients in this group of 2219 patients received their second dose in the 30th hour (data not shown). These findings suggest that the number of excluded instances of noncompliant administration of antibiotics with 24-h recommended frequencies is likely to be too low to rea- sonably alter the significance of the related findings here. Further, these findings are congruent with prior literature also demonstrat- ing an association between short recommended dosing interval and increased risk of delayed second dose administration [10].

It is also important to note that our determination of compliance

was derived using rules which, while vetted and generally endorsed by our institution, may not have been universally applied in real- time, as customization of antibiotic recommendations is not uncom- mon in clinical practice. We attempted to minimize this risk by selecting a subset of antibiotics and indications where the dosing guidelines are most commonly standardized, and by excluding pa- tient characteristics that are known to introduce variability in clini- cal practice. A notable result of this effort was the exclusion of vancomycin, which is a commonly administered antibiotic across our health system and nationally.

1. Conclusions

We found a robust statistically significant relationship between a delay in second antibiotic dose administration and increased mortality among ED patients with a broad array of bacterial infections. Significant predictors of delay included boarding status, high acuity presentation, Multiple comorbidities and use of antibiotics with short dosing fre- quency. Future Quality improvement interventions should target antibi- otic administration processes, both in the ED and during transition to inpatient care, to improve outcomes within the hospital setting.

Declarations of Interest

None.

Acknowledgements

We wish to thank Lyudmila Garbovsky, PharmD, BCPS, BCCCP and David Gajdosik, PharmD, BCPS for assistance with antibiotic Dosing standards and conceptual validation for this project.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi. org/10.1016/j.ajem.2021.02.058.

References

1. Seymour CW, Gesten F, Prescott HC, Friedrich ME, Iwashyna TJ, Phillips GS, et al. Time to treatment and mortality during mandated emergency care for sepsis. N Engl J Med. 2017;376:2235-44.
2. Rivers E, Nguyen B, Havstad S, Ressler J, Muzzin A, Knoblich B, et al. Early goal- directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med. 2001;345:1368-77.
3. Lin MY, Weinstein RA, Hota B. Delay of active antimicrobial therapy and mortality among patients with bacteremia: impact of severe neutropenia. Antimicrob Agents Chemother. 2008;52:3188-94.
4. Daniels LM, Durani U, Barreto JN, O’Horo JC, Siddiqui MA, Park JG, et al. Impact of time to antibiotic on hospital stay, intensive care unit admission, and mortality in fe- brile neutropenia. Support Care Cancer. 2019;27:4171-7.
5. Corl KA, Zeba F, Caffrey AR, Hermenau M, Lopes V, Phillips G, et al. Delay in antibiotic administration is associated with mortality among septic shock patients with Staph- ylococcus aureus bacteremia. Crit Care Med. 2020;48:525-32.
6. Brouwer MC, van de Beek D. Management of bacterial central nervous system infec- tions. Handb Clin Neurol. 2017;140:349-64.
7. Levison ME, Levison JH. Pharmacokinetics and pharmacodynamics of antibacterial agents. Infect Dis Clin North Am. 2009;23:791-815 [vii].
8. McKenzie C. Antibiotic dosing in critical illness. J Antimicrob Chemother 2011;66 Suppl 2:ii25-31.
9. Scaglione F, Paraboni L. Pharmacokinetics/pharmacodynamics of antibacterials in the intensive care unit: setting appropriate dosing regimens. Int J Antimicrob Agents. 2008;32:294-301.
10. Leisman D, Huang V, Zhou Q, Gribben J, Bianculli A, Bernshteyn M, et al. Delayed sec- ond dose antibiotics for patients admitted from the emergency department with sepsis: prevalence, risk factors, and outcomes. Crit Care Med. 2017;45:956-65.
11. Andersson M, Ostholm-Balkhed A, Fredrikson M, Holmbom M, Hallgren A, Berg S, et al. Delay of appropriate antibiotic treatment is associated with high mortality in patients with community-onset sepsis in a Swedish setting. Eur J Clin Microbiol In- fect Dis. 2019;38:1223-34.
12. Hites M, Taccone F. Dosing in obese critically ill patients. In: Udy A, Roberts J, Lipman J, editors. Antibiotic Pharmacokinetic/Pharmacodynamic Considerations in the Crit- ically Ill. Singapore: Adis; 2018. p. 47-72.
13. Blot SI, Pea F, Lipman J. The effect of pathophysiology on pharmacokinetics in the critically ill patient-concepts appraised by the example of antimicrobial agents. Adv Drug Deliv Rev. 2014;77:3-11.
14. Charlson M, Szatrowski TP, Peterson J, Gold J. Validation of a combined comorbidity index. J Clin Epidemiol. 1994;47:1245-51.
15. Joint Commission on Accreditation of Healthcare organizations. Approved: Stan- dards revisions addressing patient flow through the emergency department. Jt Comm Perspect, 32. ; 2012. p. 3-5 1.
16. Centers for Disease Control and Prevention. Core Elements of Hospital Antib- iotic Stewardship Programs. https://www.cdc.gov/antibiotic-use/core-elements/ hospital.html. [accessed 30 Nov 2020].
17. Society for Healthcare Epidemiology of America, Infectious Diseases Society of America; Pediatric Infectious Diseases Society. Policy statement on antimicrobial stewardship by the Society for Healthcare Epidemiology of America (SHEA), the In- fectious Diseases Society of America (IDSA), and the Pediatric Infectious Diseases So- ciety (PIDS). Infect Control Hosp Epidemiol. 2012;33:322-7.
18. Rhodes A, Evans LE, Alhazzani W, Levy MM, Antonelli M, Ferrer R, et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock: 2016. Crit Care Med. 2017;45:486-552.
19. R Core Team. R: A Language and Environment for Statistical Computing. https:// www.R-project.org/; 2020. Updated 2020.
20. Centers for Medicare & Medicaid Services. CMS Manual System: Pub. 100-07 State Operations Provider Certification. Revised Appendix A, Interpretive Guidelines for Hospitals. https://www.cms.gov/regulations-and-guidance/guidance/transmittals/ downloads/r77soma.pdf; Dec 22, 2011.
21. Institute for Safe Medication Practices. Guidelines for Timely Administration of Scheduled Medications (Acute). https://www.ismp.org/guidelines/timely- administration-scheduled-medications-acute; Jan 12, 2011. [accessed 15 Feb 2021].
22. Singer AJ, Thode HC, Viccellio P, Pines JM. The association between length of emer- gency department boarding and mortality. Acad Emerg Med. 2011;18:1324-9.
23. Johnson KD, Winkelman C. The effect of emergency department crowding on patient outcomes: a literature review. Adv Emerg Nurs J. 2011;33:39-54.
24. Derose SF, Gabayan GZ, Chiu VY, Yiu SC, Sun BC. Emergency department crowding predicts admission length-of-stay but not mortality in a large health system. Med Care. 2014;52:602-11.
25. Chalfin DB, Trzeciak S, Likourezos A, Baumann BM, Dellinger RP. Impact of delayed transfer of critically ill patients from the emergency department to the intensive care unit. Crit Care Med. 2007;35:1477-83.
26. Coil CJ, Flood JD, Belyeu BM, Young P, Kaji AH, Lewis RJ. The effect of emergency de- partment boarding on order completion. Ann Emerg Med. 2016;67 730-736.e2.
27. Cha WC, Cho JS, Shin SD, Lee EJ, Ro YS. The impact of prolonged boarding of success- fully resuscitated out-of-hospital cardiac arrest patients on survival-to-discharge rates. Resuscitation. 2015;90:25-9.
28. Starmer AJ, Spector ND, Srivastava R, West DC, Rosenbluth G, Allen AD, et al. Changes in Medical errors after implementation of a handoff program. N Engl J Med. 2014;371:1803-12.
29. Heilman JA, Flanigan M, Nelson A, Johnson T, Yarris LM. Adapting the I-PASS handoff program for emergency department inter-shift handoffs. West J Emerg Med. 2016; 17:756-61.
30. Starmer AJ, Schnock KO, Lyons A, Hehn RS, Graham DA, Keohane C, et al. Effects of the I-PASS nursing handoff bundle on communication quality and workflow. BMJ Qual Saf. 2017;26:949-57.